Growth OS
What actually worked, and what cannot be traced
Every analytics product will tell you what worked. The difference is what it does when it does not know. Xyranik’s answer is to say so — in a dash with a reason rather than a zero, and in a coverage figure before any share is quoted.
14-day evaluation · no card · start before you connect anything
How it works
Where this sits in the loop
Measure. It counts the real result. It follows Execute and hands on to Learn.
- 1
Each order is attributed once, in a fixed order
A discount code if there is one, then last touch, then first touch, then an explicit "we cannot tell". Attribution is written once and not recomputed later under a different model, so last month’s report still says what it said.
- 2
Coverage comes before any share
Before a channel’s share is shown, you are told how many of your orders could be traced to anything at all. A share is a share of what was traceable, and describing it otherwise would flatter every channel equally.
- 3
Incrementality comes from hold-backs, not from models
Where a campaign was sent with a deliberately excluded group, the difference between them is what the campaign added. That is the only source of an incremental figure here. Everything else is attribution, and is labelled as such.
- 4
Finished months freeze
A closed month becomes a statement that is not revised afterwards, so a report you shared with somebody does not quietly change behind you.
What you get
Attributed and incremental are never conflated
They lead to opposite decisions, so they are shown as different things throughout. A business acting on one while thinking it has the other is the specific mistake this prevents.
Return on investment as a ladder
Every sale, then sales from people the marketing reached, then sales the marketing gets credit for, then sales it can be shown to have caused. Each rung is a narrower claim than the one above it, and the narrowing is visible.
Comparisons refuse when the basis moved
If attribution coverage shifted materially between two periods, the comparison is refused rather than presented — because the change in the number would mostly be a change in what could be seen.
What it does not do
Listed here rather than discovered later.
Incrementality needs a hold-back to exist
Without a campaign that deliberately excluded a group, there is no incremental figure — and Xyranik will show a dash rather than infer one.
Platform engagement figures are not imported
Likes, reach and impressions are not read into Xyranik from the social platforms, so they are not shown. What is counted is clicks on tracked links and the orders that followed.
Give it a goal and watch what it proposes
Nothing reaches a customer or spends money without you. The evaluation starts before you connect anything, so you can read what it would do before deciding whether it may.